Decreases in Colwell bicarbonate soil test P in the years after addition of superphosphate, and the residual value of superphosphate measured using plant yield and soil test P

Decreases in Colwell bicarbonate soil test P in the years after applying si ngle (ordinary) superphosphate, and the residual value of superphosphate, w as measured in a long-term field experiment on a duplex (texture contrast) soil (sand over lateritic ironstone gravel clay sand at 10-15 cm), at Wonga n Hills, Western Australia, typical of many soils used to grow crops in Wes tern Australia. Ten levels of P (0-91 kg P ha(-1)) were applied once only i n late May to different plots in different years from 1988 to 1993. Wheat ( Triticum aestivum), or lupin (Lupinus angustifolius)) were sown in late May of each year, when the P treatments applied that year were banded (drilled ) with the seed. Soil samples were collected each June to measure soil test P. Seed (grain) yields of the crops were measured each December. The resid ual value (RV) of P applied in previous years was calculated relative to P applied in the current year, using grain yields (RVyield) and soil test P ( RVsoil). Soil test P measured on soil samples collected in June was related to yields measured in December that year to provide soil P test calibratio ns. Relative to P applied in the current year, soil test P decreased by bet ween 15 to 30% for P applied one year previously, by 25 to 30% for P applie d three years previously, and by 60 to 70% for P applied six years previous ly. Soil test P was affected by spatial variation, and it also varied in th e different years, for P applied in the current year, one year previously, two years previously, etc. Compared with P applied in the current year, mea n RVyield determined in the different years decreased by about 40% one year after P application, followed by a further 20% decrease for P applied two years previously, followed by a further 20% decrease for P applied three to five years previously. Relative to current P, RVsoil decreased by about 25 % one year after P application, followed by a further 20% for P applied two years previously, followed by a further 10% for P applied three years ago, and followed by a further 6% for P applied four and five years ago. As mea sured in the different years, the soil P test calibration varied between ye ars for P applied one, two etc. years previously. This was so even when the same cultivar of wheat was grown at the same site in different years.